Database/System/Theory

• Database Systems Theory

  • Overview

    • Database systems theory studies how a Data/Base represents information, answers queries, and keeps data usable when many operations occur or a machine fails. The data model describes what can be stored; the query model describes what can be asked; the storage engine determines how those operations run.

  • Core questions

    • Modeling — A Database/Schema names the structures and constraints that data must satisfy. The logical model describes the data independently of indexes and file layout.
    • Querying — A query describes the desired result. Indexes and query plans determine how the system finds that result efficiently.
    • Transactions — Concurrency control keeps simultaneous operations from producing invalid outcomes; recovery restores a consistent state after failure. Durability, isolation, and consistency guarantees depend on the system's design.
    • Distribution — Replication and partitioning spread data across machines, introducing choices about coordination, availability, and when replicas see changes.

  • Data models and workloads

    • Relational databases organize data as relations with keys and constraints; SQLite and Dolt are examples. SQL includes DDL for defining structures and statements for querying and changing rows.
    • Graph/Database represents relationships as edges, making connected paths central to the model.
    • Data/Base/Vector indexes embeddings for similarity search; vector search can also be an extension of another database.
    • Document and key-value databases organize records around documents or keys. ArangoDB combines these models with a graph model; schemas and query capabilities vary by product.
    • Analytical databases favor scans and aggregation across large datasets; OLAP describes this style of work, while transactional systems favor short reads and writes.